Setup guide

Set up Nexus from install to first search.

This guide walks through activation, AI setup, job-site sign-ins, resume setup, and your first search so you can get started without guesswork.

Install
Activate
Connect AI
Run first search

Fast path

1. Activate NexusInstall the app and apply your license key first.
2. Sign in + test setupVerify your AI connection and sign in to the job sites you plan to use.
3. Start with one small searchRun a low-volume first pass before increasing daily applications.
Manage License Request Access

1. Preflight Checklist

Before you get started, make sure these five things are ready:

  • You have your license key in format NEXUS-XXXX-XXXX-XXXX-XXXX.
  • Your target platforms are signed in at least once manually (LinkedIn and/or Indeed).
  • Your resume PDF is finalized and saved locally.
  • If using the built-in AI, 6GB+ free RAM is recommended for lightweight models.
  • You know your daily application limits for each platform.
Goal: get one successful end-to-end application first, then scale volume.

2. Install + Activation

  1. Download the latest Nexus build from the beta channel.
  2. Install and launch the app.
  3. Paste your license key on the activation screen and click Activate.
  4. If activation fails: verify internet connectivity and retry once.
Once a machine is activated, you usually will not need to enter the key again unless you move devices.
Nexus activation screen requesting a license key
Activation happens before the full workspace opens. Once the key verifies, Nexus unlocks the full desktop app.

3. AI Setup

Option A: Built-in AI (Recommended)

Nexus includes a built-in AI runtime. This is the default and fastest path for new users.

  1. Open Settings → LLM Settings.
  2. Select provider: Nexus Built-in (Recommended).
  3. In the Model Manager section, click Download on your preferred model:
    • phi-3-mini (~2.3 GB) for lightweight systems.
    • qwen2.5-7b (~4.7 GB) as the recommended text baseline.
    • qwen2-vl-2b-visual (~2.4 GB) for external apply recovery.
  4. Once downloaded, click Start to launch the local server.
  5. Click Test Connection to verify.
Built-in text + visual runtimes can auto-start on demand once models are downloaded.

Option B: Ollama / LM Studio

  1. Use this only if you specifically prefer your own local runtime.
  2. Install Ollama or LM Studio from their official sites.
  3. Download qwen2.5-7b-instruct (or your preferred model) and start the local server.
  4. In Nexus Settings, switch provider to Ollama or LM Studio.
  5. Use http://localhost:11434/v1 for Ollama or http://localhost:1234/v1 for LM Studio.

Option C: OpenAI / Anthropic

  • Set provider to OpenAI or Anthropic.
  • Enter your API key. Job descriptions will be sent to the cloud provider.

Verify connection

  1. Click Test LLM Connection in Settings.
  2. Expected result: Connected! Model: ...
  3. If it fails, read the returned endpoint and status message carefully.

4. Job Site Sign-ins

Use the Accounts page in Nexus to verify platform sessions before running searches.

  1. Open the LinkedIn session window and sign in.
  2. Open the Indeed session window and sign in.
  3. Refresh status inside Nexus until each shows connected.
If Nexus says a site is not signed in, refresh that session once from the Accounts page before starting again.
Nexus Accounts page showing job-site session controls
Keep at least one active session per platform you plan to automate before launching a search.

5. Resumes + Profile Defaults

  1. Upload your resume in the Resumes tab.
  2. Confirm the file path via “Open location” when needed.
  3. Set defaults for location, work authorization, phone, and work links.
  4. If available in your build, set demographic defaults carefully.

For consistent outcomes, keep one primary resume per search and avoid switching mid-run.

Nexus Resumes page with multiple uploaded resumes available for searches
Upload your core resume set once, then attach the right file to each search instead of swapping files mid-run.

6. Create Your First Search

Baseline search recipe

  1. Choose site (LinkedIn or Indeed first).
  2. Add keyword + location.
  3. Select a resume for this search explicitly.
  4. Set minimum match score (for example: 7.5 / 10).
  5. Set a daily application limit (for example: 10 for calibration).
  6. Set Search Style to Balanced (Recommended).

Good first-pass defaults

  • Start with required fields plus one resume.
  • Use Balanced Search Style for most first runs.
  • Leave advanced options collapsed until the first stable run.
  • Keep the daily limit low until the first run is stable.
Nexus Create Search flow with recommended defaults for a first search
For the first pass, use one resume, Balanced Search Style, and a modest daily limit until the workflow proves itself.

7. Watch the First Run

  1. Start one search and watch the first 2 to 3 applications closely.
  2. Confirm question-answering behavior across text, dropdown, and radio fields.
  3. Check activity logs for applied, skipped, and failed counters.
  4. Review the post-run summary and adjust thresholds.

After one stable run, scale to more pages and higher daily limits.

8. Troubleshooting

LLM test shows “connection failed”

  • If using Nexus Built-in: ensure a model is downloaded and the server is started.
  • If using Ollama: verify the server is running and the model is loaded. Use http://localhost:11434/v1.
  • If using LM Studio: verify the server is running and a model is loaded. Use http://localhost:1234/v1.
  • Leave the model field blank for auto-select or use the exact model ID.

Captcha interruptions

  • Invisible reCAPTCHA: sign in to the trusted account and retry.
  • hCaptcha: complete the challenge manually when prompted.

Easy Apply rate limits

  • If rate-limited and the search is Easy Apply only, stop it and retry the next day.
  • If external apply is enabled, skip easy applies and continue with supported external flows.

Form question validation failures

  • For numeric-only fields like GPA, avoid text responses.
  • For required dropdown or radio fields, ensure defaults include location and authorization context.

9. Scale Up Safely

Once your first search is clean:

  • Run small batches, validate quality, then scale volume.
  • Keep one known-good config profile as a fallback baseline.
  • Log unsupported domains and prioritize modules by frequency.
  • Rotate between high-quality keywords instead of broad scraping.
  • Review post-run outcomes daily and tune thresholds weekly.
Quality beats quantity. Higher acceptance and fewer platform flags come from controlled pacing.

10. Get Help

If you need help with a run, include these artifacts in your report:

  • Search ID and timestamp window
  • Relevant search logs
  • Target URL and platform
  • Screenshot of the blocking UI state

Contact support at Hyperfectllc@gmail.com and include “Nexus Beta Support” in the subject line.